R&D cooperation has received great attention among industrialists, decision-makers and researchers. This article introduces an econometric approach for identifying the factors that lead firms to cooperate in order to achieve innovation. The determining factors were defined according to empirical findings from the Korean Innovation Survey (KIS) 2005, captured in a multivariate probit regression model. The aim is to subsidize further research by applying agent-based modelling to simulate innovation networks in the Korean manufacturing sector. ©2012 Taylor & Francis.